Strategije u učenju trećeg jezika

1 Sveučilište u Zagrebu Filozofski fakultet Odsjek za anglistiku Katedra za metodiku Strategije u učenju trećeg jezika Diplomski rad Student: Marija Novokmet Mentor: dr. sc. Stela Letica Krevelj Zagreb, studeni 2015.

2 University of Zagreb Faculty of Humanities and Social Sciences Department of English TEFL Section Language learning strategies in third language acquisition Graduation Thesis Student: Marija Novokmet Supervisor: Stela Letica Krevelj, Ph.D. Zagreb, November 2015

3 Examining Committee: Assistant Professor Renata Geld Anđel Starčević, Ph.D., postdoc. Stela Letica Krevelj, Ph.D., postdoc.

4 Table of Contents 1 Introduction... 6 2 Language learning strategies and their taxonomies... 7 3 Third language acquisition... 9 4 Multilingualism... 11 5 Studies on strategies in third language acquisition... 15 6 Study... 19 6.1 Aims... 19 6.2 Participants... 19 6.3 Data collection instrument... 21 6.4 Data collection procedures... 22 6.5 Results and discussion... 22 6.5.1 Language learning strategies questionnaire analysis... 22 6.5.1.1 Analysis of the items taken from Oxford's SILL (1990).....24 6.5.1.2 Analysis of items which tested cross-linguistic strategies... 27 6.5.2 The analysis of open-ended questions... 30 7 Conclusion... 34 8 References... 36 9 Appendix... 43

5 Abstract Language learning strategies are specific actions that learners employ in order to make their learning easier, faster, more self-directed and efficient (Oxford, 1990). It is of great interest to both teachers and students to know what these techniques are and how they can be employed, in order to improve their teaching and learning. Even though there is evidence that bilingualism and multilingualism may positively affect an additional language learning in various ways, some studies (Kemp, 2007, Korkmaz, 2013, Mitits & Sarafianou, 2012, Psaltou-Joycey & Kantaridou, 2009, Qasimnejad & Hemmati, 2013, Sung, 2011) have been undertaken to find how exactly multilinguals apply their vast linguistic knowledge when approaching various linguistic tasks in terms of strategies used. This paper investigated the use and frequency of language learning strategies of two multilingual groups of Croatian high school students who had been studying two or three foreign languages, respectively. The research study was carried out using statments based on items from Oxford's SILL questionnaire (1990), and additional items, which were added with the aim to test crosslinguistic strategies in particular. The results showed that the participants with the knowledge of three languages used a wider variety of strategies and used them more often. The same was found in the case of cross-linguistic strategies. Key words: language learning strategies, third language acquisition, bilingualism, multilingualism

6 1 Introduction In trying to discover what makes good language learners, the issue of language learning strategies was raised in order to see whether the use of strategies makes them different from other, less successful learners. Among factors related to choice of language learning strategies are language being learned, level of language learning, degree of metacognitive awareness, gender, motivation and language language learning goals (Oxford & Nyikos, 1989). The studies in third language acquisition field have shown that bilinguals had some advantages over monolinguals in learning additional languages, for example, they had heightened levels of metalinguistic and crosslinguistic awareness which play an important role in the development of their language learning strategies (Aronin & Singleton, 2012, Jessner, 2006, Cenoz & Jessner, 2009). The aim of this paper is to investigate the influence of the number of languages known on language learning strategy use and the use of crosslinguistic strategies. The first part of the paper will present the theoretical background of language learning strategies and an overview of the field of third language acquisition. The definitions of the term "language learning strategies" will be presented, along with the comparison of the major taxonomies of language learning strategies. This part also deals with third language acquisition and gives explanations of important terms related to it, such as bilingualism, multilingualism, metalinguistic awareness and crosslinguistic interaction. Finally, this part will provide a review of some studies that investigated language learning strategies of multilingual learners. In the second part, the study carried out with two groups of multilingual learners showing the differences in the choice and frequency of language learning strategies used by learners who learn two and those who learn three foreign languages will be presented.

7 2 Language learning strategies and their taxonomies When it comes to learning, some individuals seem to perform better than others who struggle with acquiring new knowledge. Those differences can also be compared to learning new languages, which is an especially complex and long-lasting undertaking. Some of the differences between effective, fast and ineffective, slow language learners can be attributed to special techniques they use when learning a new language, that is, language learning strategies. Along with a growing interest in what makes a ''good'' language learner in recent decades, many different definitions in second language acquisition field of what learning strategies are have emerged. However, as there is still no agreement on what exactly they are and how they should be classified, a definition of language learning strategies that would be unanimously accepted has yet to be created. Firts of all, as there is not a consensus on the terminology, different authors have used different terms, such as "learner strategies" (Rubin, 1987), "learning strategies" (O'Malley & Chamot, 1990) and "language learning strategies" (Oxford, 1990). Dörnyei (2005, 2006, as cited in Cohen, 2011), on the other hand, proposed the term "learner self-regulation", which referred to the degree to which individuals were active participants in their own learning, arguing that the process of self-regulation merely generated strategies as a product. Oxford defined language learning strategies as "specific actions taken by the learner to make learning easier, faster, more enjoyable, more self-directed, more effective, and more trasferrable to new situations'' (1990, p. 8). She claimed that strategies were especially important for language learning because they were ''tools for active, self-directed involvement, which was essential for developing communicative competence'' (Oxford, 1990, p. 1). O'Malley and Chamot (1999) defined language learning strategies from a cognitive theory point of view and claimed they were ''special ways of processing information that enhance comprehension, learning or retention of the information'' (p. 1). The expectation based on their view would be that strategies are located in the brain and that learning a language does not differ from learning content, and therefore does not have a role for any specific linguistic faculty (O'Malley & Chamot, 1990, as cited in Skehan, 1991). However, Chamot and El-Dinary (2000, as cited in Macaro, 2006) proposed that strategies were not only mental procedures that assisted learning but that they also included overt activities. Similarly, Oxford's definition (1990) included physical activities, such as writing in a notebook, or physically acting out new words, as examples of strategic behaviour. Phakiti

8 (2003, as cited in Macaro, 2006) defined strategies not in the strictest sense of the term and argued they should be seen as learners' stable long-term knowledge of their strategy use. He, therefore, ruled out action as an integral component of a strategy. Mayer (1988, as cited in Macaro, 2006) referred to strategies as "behaviours of a learner that are intended to influence how the learner processes information" (p. 11). In his vew, which is similar to Oxford's (1990), strategies were additional to the processing of information in a sense that they facilitated learning and made it more effective (Macaro, 2006). In a similar vein, White (2008) defined language learning strategies as the operations or processes which were consciously selected and employed by the learner to learn the target language or facilitate a language task. Brown (2000) stated that language learning strategies were moment-by-moment techniques that we employed to solve problems posed by second language input and output. When the learner consciously chooses strategies that fit his or her learning style and the L2 task at hand, these strategies become a useful toolkit for active, conscious, and purposeful self-regulation of learning (Oxford, 2003). According to Cohen (2011), language learning strategies could be defined as thoughts and actions, consciously selected by learners, to assist them in learning and using language in general as well as in the completion of specific language tasks. Cohen made a distinction between language learning strategies (i.e., strategies for the learning of language material for the first time) and language use strategies (i.e., strategies for using the material that has already been learned to some degree), which are deployed in complex and interacting ways, so that at any given moment it is difficult to isolate a single strategy being used. Griffiths (2013) identified six essential characteristics of language learning strategies: they are active, conscious, chosen, purposeful, regulatory and learning-focused. However, Griffiths (2013) argued that ''deliberate'' versus ''automatic'' was a more useful distinction than ''conscious'' versus ''unconscious'', and that strategies could operate somewhere on a continuum between the two. Furthermore, Bialystok (1978, as cited in Griffiths, 2013) explained that language learning strategies were optional means for exploiting available information to improve competence in a second language. It is then, logical, that strategies have to be chosen by learners, since it would be impossible to force them to employ them against their will. According to Griffiths (2013), learning activities that are passively accepted from others can hardly be called strategic. Furthermore, strategic activity must be

9 purposefully related to a goal, and not just some kind of random behaviour, and it has to be aimed at the regulation of language development (Griffiths, 2013). Some of the most influential classifications of language learning strategies that have appeared over the years were those of O'Malley, Chamot, Stewner-Manzanares, Küpper and Russo (1985), Oxford (1990) and Rubin (1987). All three of them distinguished between different categories of strategies depending on the level or type of processing involved. O'Malley et al. (1985, as cited in O'Malley and Chamot, 1999) differentiated between metacognitive (higher order executive skills that refer to planning for, monitoring, or evaluating the success of a learning activity), cognitive (strategies that operate directly on incoming information and manipulate it in ways that enhance learning and are mostly used for the specific type of task) and social/affective strategies (strategies that entail either interaction with another person or control over affect). Both Rubin (1987, as cited in Zare, 2012) and Oxford (1990) distunguished between strategies contributing directly to learning and those contributing indirectly to learning. Rubin further classified direct strategies into metacognitive and cognitive, and indirect into communicative and social strategies. Oxford subdivided these two classes into a total of six groups. Direct strategies consist of memory, cognitive and compensation strategies, while indirect strategies include metacognitive, affective and social strategies. The compensation category seems to be a major addition in Oxford's classification. As can be seen, in defining language learning strategies the focus has shifted over the years from their product (linguistic or sociolinguistic competence) to their processes and characteristics (Lessard-Clouston, 1997). However, there are many problems in language learning strategy research, which can be summarised as follows: there is no consensus on what language learning strategies do, especially whether they are always facilitative and effective, whether they occur inside or outside the brain, whether they consist of knowledge, intention, action or all three, whether their integrity survives across all learning situations, tasks and contexts, how general or abstract they are and whether and how they could be classified in a framework or a hierarchy (Macaro, 2006). 3 Third language acquisition In recent times, a significant amount of research on third language acquisition and multilingualism has appeared. Consequently, there has been a lot of effort by experts in the

10 area of language learning to give a precise definition of a third language (L3) and establish its difference in relation to the first (L1) and the second language (L2). The term L3 has been used variably in the literature since it is not self-evident what kind of concept it can refer to. Hammarberg (2009) explained that one common practice was to number the speaker's languages chronologically, according to the time of first encounter: L1, L2, L3, L4 and so forth. This chronological scale may seem parallel to the terms monolingual, bilingual, trilingual, quadrilingual and so forth, which represent the result of the acquisition of a certain number of languages. According to Hammarberg (2009), the analogy is superficial, since it will often not be possible to order a multilingual's languages along a linear time scale. Therefore, Hammarberg proposed using the term third language (L3) for a non-native language which is currently being used or acquired in a situation where the person already has a knowledge of one or more L2s (languages encountered and acquired after infancy) besides one or more L1s (languages acquired during infancy). ''An L3 is thus a special case of the wider category of L2, and not necessarily language number three in order of acquisition'' (Hammarberg, 2009, p. 6). Moreover, De Angelis (2007) suggested the idea of using L3 to refer to a third or additional language, regardless of whether it is a third, fourth or sixth language. Although many different terms have been put forward to name the field itself, such as Multiple Language Acquisition, Multilingual Acquisition and Third Language Acquisition (TLA), De Angelis (2007) argued that the term Third or Additional Language Acquisition was the most proper one because it refered to all languages beyond L2 without giving preference to any particular language. Most scholars working with L3 learners agree that learning an L3 differs from learning an L2 in many respects and that a clear distinction between Second Language Acquisition (SLA) and Third Language Acquisition (TLA) has to be drawn. Jessner (1999) argued that SLA had to be treated differently from TLA due to the increasing complexity of crosslinguistic interaction, the importance of metalinguistic awareness in the acquisition process of an L3, and the increased pressure from language attrition and relearning. One of the most interesting issues regarding TLA is to see whether bilinguals have advantages over monolinguals in learning a further language and to look into the effects of bilingualism on TLA. The majority of studies on general proficiency indicated a positive effect of bilingualism on TLA which was also linked to language learning strategies and communicative ability, particularly in the case of typologically close languages (Cenoz & Jessner, 2009).

11 Consequently, when faced with a complex task of acquiring L3, bilingual students are able to use a wider variety of learning strategies and to weigh the effectiveness of these strategies due to their experience in learning languages (McLaughlin & Nayak, 1989, as cited in Molnár, 2008). Prior language learning experience changes the quality of L3 learning which results in differing language strategies which the experienced language learner develops in contrast to the inexperienced one. These advanced cognitive skills in language learning may, therefore, lead to the speeding up of the language learning process (Jessner, 1999). The complexity of TLA is also linked to individual factors in language learning. The interplay between the various individual factors, such as aptitude, motivation, learning strategies, learning styles and L1 influencing the second language learning process is rather complex so it is not surprising that in TLA the complexity increases (Jessner, 2008). In other words, it is difficult to determine the extent to which each of these factors contribute to the success in learning an L2 and that difficulty only increases with an additional language in a learner's repertoire when learning an L3. For instance, the interplay between an L1 and L2 makes learning an L3 more complex because the learner has more linguistic information he can draw conclusions from. The number and variety of learning strategies increases as well. 4 Multilingualism As Jessner (2008) stated, bilingualism refers to the mastery of two languages, while multilingualism refers to the familiarity with more than two languages. Multilingual acquisition is a complex and dynamic process which can be a result of either the simultaneous or the consecutive acquisition of foreign languages. When two languages are involved there are only two possibilities: early bilingualism when the two languages are learned simultaneously and second language acquisition when they are learned consecutively. In third language acquisition there are at least four possibilites: A. Simultaneous acquisition of L1/L2/L3; B. Consecutive acquisition of L1, L2 and L3; C. Simultaneous acquisition of L2/L3 after learning the L1; and D. Simultaneous acquisition of L1/L2 before learning the L3 (Cenoz, 2000, as cited in Cenoz & Jessner, 2009).

12 Multilinguals seem to possess special characteristics not found in monolinguals or bilinguals. Relative to monolinguals or bilinguals they have larger overall linguistic repertoires, tend to use more learning strategies, seem to have enhanced metalinguistic awareness, seem to acquire greater sensitivity to socio-pragmatic aspects of communication, and are more responsive to both linguistic and non-linguistic factors in communicative situations (Aronin & Singleton, 2012). Language learning skills and language maintenance skills are developed at a higher level in multilinguals as well. They are also more adept at language managment, which is the multilingual art of balancing communicative requirements with language resources. These skills seem to contribute to metalinguistic or language awareness (Cenoz & Jessner, 2000). Jessner (2006) put forward the idea of linguistic awareness in multilinguals and defined it as an emergent property of multilingual proficiency consisting of at least two dimensions in the form of crosslinguistic awareness and metalinguistic awareness. Crosslinguistic awareness refers to the learner s tacit and explicit awareness of the links between their language systems, while metalinguistic awareness refers to the ability to focus attention on language as an object in itself and to think abstractly about language, to play with or manipulate language. This is multilinguals' most characteristic cognitive ability (Jessner, 2006). Yopp (1988, as cited in Thomas, 1992) reported four general types of metalinguistic ability: phonological awareness, word awareness, syntactic awareness, and pragmatic awareness. According to Mora (2001, as cited in Szerencsi, 2010), metalinguistic awareness incorporates the skills of knowing that language has a potential greater than that of simple symbols (it goes beyond the meaning); that words are separable from their referents (meaning resides in the mind, not in the name), and that language has a structure that can be manipulated (language is malleable). Metalinguistic awareness allows the individual to step back from the comprehension or production of an utterance in order to consider the linguistic form and structure underlying the meaning of the utterance. Metalinguistic awareness refers to knowing how to approach and solve certain types of problems which themselves demand certain cognitive and linguistic skills (Malakoff, 1992, as cited in Jessner, 2006). In Bialystok's view (2001, as cited in Kuo & Anderson, 2008), metalinguistic awareness involves both the analytical ability to reflect upon and manipulate formal properties of language and the attentional control of the mental mechanism that operates language processing. Bialystok (1978, as cited in Harris, 1992) made a distinction between explicit and implicit knowledge and proposed that facts stored as explicit linguistic knowledge could be articulated in contrast to automatic information that

13 was used spontaneously and was respresented in implicit linguistic knowledge. Implicit linguistic knowledge contains the information about the target language necessary for spontaneous comprehension, that is, it functions communicatively. Whereas Bialystok emphasized non-communicative functions of explicit knowledge, Odlin (1986, as cited in Harris, 1992) stressed communicative functions of metalinguistic knowledge, suggesting that bilinguals who code-switch were often aware of their lexical choices and, therefore, used formal linguistic knowledge with a communicative purpose. As can be expected, metalinguistic and metacognitive awareness play an important role in the development of language learning strategies in multilingual learners and users (Cenoz & Jessner, 2009). Furthermore, metalinguistic awareness is closely connected with language aptitude. The bigger the number of languages involved in the acquisition process, the more difficult it is to decide whether language aptitude or metalinguistic awareness influence the language acquisition progress (Jessner, 2006). The role of an increased level of metalinguistic awareness, which forms an integral part of multilingual proficiency, was emphasized in the Dynamic Model of Multilingualism developed by Herdina and Jessner (2002, as cited in Marx & Hufeisen, 2004). This model accounts for linguistic development which involves the cumulative interdependence between language systems of a multilingual (i.e. L1, L2, L3, etc.), resulting in complex cross-linguistic interactions and development of specific skills tied to language learning (Herdina & Jessner, 2002, as cited in Marx & Hufeisen, 2004). Role-function model by Hammarberg and Williams (2001a, 2001b, as cited in Marx & Hufeisen, 2004) illustrated different roles the various background languages of a learner might fulfil in the spoken production of the target language. The model distinguished between the instrumental language and the default supplier language. The instrumental language is metalinguistic in nature and performs an instrumental function for target language production, while the default supplier language is the main source for crosslinguistic influence, slipping into target language production without the learner's intention. Only one language will be chosen as the default supplier and the choice depends on how well each of the languages in an individual's repertoire fulfils four criteria: typological similarity (how related the languages are to each other or how related the learner believes they are), proficiency (how well the languages are spoken), recency of use (how often the languages are spoken), and its status as an L2 (foreign languages are more likely to become default suppliers than L1s, while L1s are usually chosen as instrumental languages). Possible reasons for the greater possibility of using

14 an L2 instead of an L1 as the default supplier are different acquisition mechanisms for foreign languages as opposed to the L1, which leads to an activation of the L2 type of mechanism in L3 acquisition, and a desire to suppress the Ll as ''non-foreign'' and instead orientate oneself towards a prior foreign language when approaching the target language (Hammarberg, 2001a, as cited in Marx & Hufeisen, 2004). This might be due to the fact that foreign languages are generally perceived to be closer to each other than to the native language (De Angelis, 2007). It is worth noting that the roles of the languages may shift over time, with increased competence in the target language leading to a decreased reliance on various other languages. (Hammarberg, 2001a, 2001b; Williams & Hammarberg, 1998, as cited in Marx & Hufeisen, 2004). Yet another model that has been developed to explain and describe multiple language acquisition was Hufeisen s Factor Model (Marx & Hufeisen, 2007, 2004), which emphasized the differences between learning an L2 and learning an L3. Whereas at the beginning of the L2 learning process the learner is a complete novice in the learning process of a second language, in third language learning the learner already knows how to approach a new language. The learning of an L3, therefore, includes a new set of learning factors, that is, foreign/second language learning-specific factors, such as individual second language learning experiences, interlanguages of other learned languages, and foreign language learning strategies. Each factor interacts with the others, but the importance of single factors and their relevance for the success of the learning process is different from learner to learner. This model does not assume difference in learning an L3 and additional languages beyond it (Marx & Hufeisen, 2007, 2004). The Multilingual Processing Model (Meißner, 2003, Meißner & Burk, 2001, as cited in Marx & Hufeisen, 2004) suggested that learners constructed a type of ''spontaneous grammar'' in the new language, based on what they know of other, related languages and modified according to multilingual strategies. Those strategies strengthen hypothesis formulation and sensitise the learner to differences between languages in terms of syntax or lexis. This model is similar to the role-function model in that it proposes that learners use one or more language systems that the learner already knows and which are called on when understanding the target language - providing these language systems are etymologically related with the new language and the learner had reached a certain degree of competence in them (Meißner, 1998, as cited in Marx & Hufeisen, 2004). Each new language in an individual's repertoire adds to the individual's ''plurilingual intergrammar'', which functions as

15 a base for any languages that were to be learnt in the future (Meißner, 2003, Meißner & Burk, 2001, as cited in Marx & Hufeisen, 2004). In his strategy model of multilingual learning, Müller-Lancé (2003) offered some essential aspects of the multilingual mental lexicon, such as that the connections between the elements of different foreign languages were not necessarily weaker than those between foreign language elements and L1 elements. Furthermore, he stated that mental connections had different "strengths": extremely strong were those between cognates, that is, phonetically and semantically related words of different languages. When it came to experienced learners, cognates of different languages were more strongly connected mutually than to the respective L1 element, or to the other elements of the respective foreign language. Accordingly, interlingual connections can be stronger than intralingual connections. It is also plausible that the respective forms of L1, L2, L3 etc. are grouped around one common concept. Finally, learners usually make semantic connections between the words, rather than phonetic (Müller- Lancé, 2003). 5 Studies on strategies in third language acquisition Although a lot of research has been done on multilingualism and L3 acquisition in recent years, there is still little information on how language learning strategies affect L3 acquisition. Jessner (2008) offered a summary of some studies that dealt with multilingual learner strategies. One of those studies, carried out by Naiman, Frohlich, Stern and Todesco (1996[1978]), looked into the characteristics of good language learners and found their success was linked to a number of strategies, such as an active learning approach, realization of language as a system, realization of language as a means of communication, handling of affective demands and monitoring of progress. Nayak, Hansen, Krueger and McLaughlin (1990, as cited in Cenoz & Genesee, 1998) wanted to see if multilingual learners used different strategies to learn a miniature linguistic system assigned in the study than did monolingual learners. The results showed that the multilingual learners were superior to monolinguals in a number of ways: they demonstrated greater flexibility in switching strategies according to the demand characteristics of the task for example, they preferred mnemonic strategies for a memory task and linguistic strategies

16 for a rule-discovery task; they were more likely to modify strategies that were not effective in language learning and were more effective using implicit language learning strategies. Thomas (1992, as cited in Cenoz & Genesee, 1998) found that bilinguals who were learning an L3 used more communication strategies in comparison with monolinguals who were more worried with grammar and vocabulary. The author explained that the fact that bilinguals had to switch languages according to the situational demands could enhance their sensitivity to the functions of language for social communication. It was concluded that a student s prior linguistic experience influenced the strategies which they subsequently adapted, and their success in the foreign language classroom (as cited in Jessner, 2008). Similarly, Mißler (1999, 2000) found that the increase of language learning experience was reflected in the number of strategies, which also turned out to depend on individual factors (as cited in Jessner, 2008). The results of Kemp's study (2001) showed that multilinguals developed more grammatical metalinguistic awareness and were, therefore, better at learning additional languages. In her later study (2007), she investigated the use of grammar learning strategies and found that the more languages learners knew, the greater the number of grammar learning strategies they used and the more frequently they used them. This led her to a conclusion that "there may be a threshold effect for the use of grammar learning strategies so that an increase in the number and frequency of strategies used occurs to a greater extent during the acquisition of the third language, increasing more gradually in additional languages, than occurs in L2 learning" (Kemp, 2007, p. 241). The reason behind this may be bilinguals' lack of procedural knowledge of processing that multilinguals have, or the automaticity of use, which multilinguals are able to apply in learning another language (Kemp, 2007), and which allows them to pay greater attention to important aspects of the input. Additionally, the participants in the study were asked to write in strategies they used that were not present in the questionnaire (i.e. idiosyncratic strategies). The most notable strategy that they named was "I compare and contrast the target language grammar with the grammar in other languages I know" (Kemp, 2007, p. 251). This showed that participants did not only have implicit procedural knowledge of their grammar learning strategies, but were also able to identify them (Kemp, 2007). The study showed that prior linguistic knowledge played an important part in how learners approached the study of a target language. Mitits and Sarafianou (2012) conducted a study involving three successful language learners (two bilinguals with L1 Greek-L2 Serbian and L1 Russian-L2 Greek and a

17 monolingual with L1 Greek) in order to observe how language learning strategies developed across languages and whether bilingual learners' use of strategies differed quantitatively and qualitatively from that of monolinguals when learning English L3. The results showed that the bilingual learners used more strategies more frequently than the monolingual one. There was also a qualitative difference in strategy use in that the bilinguals tended to transfer more strategies that have to do with implicit rather than explicit language learning and relied more on the typological similarities between languages. Bilingual participants reported a number of strategies they used when encountering new vocabulary, such as drawing from L1 and L2, depending on the word closeness, guessing from context or analysing word parts to come to understanding. They also reported constantly learning from their mistakes and avoiding literal translation. L1 Russian-L2 Greek participant stated that Russian often helped her with her English, but sometimes found it confusing and mixed codes, particularly when lacking the knowledge of Greek vocabulary. Qasimnejad and Hemmati (2013) also investigated the difference between monolinguals and bilinguals in terms using language learning strategies. The participants were monolingual Persian and bilingual Turkish and Persian university students as EFL learners. The results revealed that both monolingual Persian and bilingual Turkish-Persian university students employed a wide variety of language learning strategies, with bilingual learners employing more strategies than their monolingual peers. The Turkish-Persian students reported using metacognitive strategies to control learning by for example, thinking about their progress in learning English, seeking out ways to improve their learning, setting clear goals in learning English, and planning their schedule to study English. More importantly, they reported greater tendencies than monolinguals to notice their mistakes and to be able to learn from them. The most frequently used strategies among bilinguals were metacognitive, which they used much more often when compared to monolingual students. The second most frequently used strategies by bilingual students were compensation strategies, which they found useful in overcoming their missing knowledge of English through the use of synonyms, guessing, and reading English without looking up every new word. Cognitive strategies were ranked as the third amongst preferred strategies in both groups, although use was higher for the bilingual group. When it comes to affective strategies, both groups were equally aware of their anxiety and tension, however, the bilingual group reported using them more often. Social strategies were reported among the least used strategies by both bilinguals and monolinguals. Still, the majority of the respodents in both

18 groups reported always or almost always trying to practice English with other students and native speakers. Memory strategies were least favoured of the six categories of strategies by bilinguals. Sung (2011) investigated Chinese FL learners' strategy use in the US universities and came to a conclusion that the participants who had studied one other foreign language prior to Chinese used the four categories of strategies, cognitive, metacognitive, affective and social strategies, less frequently than those who had studied two or more other foreign languages prior to Chinese. In other words, the more languages the learners studied, the more frequently they used strategies in learning the additional language, Chinese. Psaltou-Joycey and Kantaridou (2009) conducted a similar study on language learning strategy use, but with bilingual and trilingual Greek university students learning FLs in an academic context. The results of the study indicated that trilingual students used more strategies more frequently than bilinguals. Trilinguals outperformed bilinguals in the use of metacognitive, cognitive, compensation and social strategies. To be more specific, trilingual students reported they made associations between new elements in the target language and what they already knew in order to memorise them and used guessing and synonyms to find out new meanings. They also reported not hesitating to take risks and proved to be more willing than bilingual students to speak in the foreign language. A study by Korkmaz (2013) explored the most and the least frequently used language learning strategies of TEFL students with Turkish L1 when learning German or French as their L3. The results revealed that the participants from both groups employed the same strategies; compensation strategies emerged as the most frequently used ones, memory strategies as the secondly most frequently used, whereas affective strategies as the least frequently used ones. Using guessing strategy and associating known subjects with new ones were two frequently used strategies for both groups. The participants also reported that they mostly used English, their L2, to guess and associate the meaning in their L3s, which indicated that language learning strategies were transferable and the strategies developed when learning a first foreign language had valuable contribution to learn a later one (Korkmaz, 2013). In a similar vein, a study by Molnár (2008) reported that knowledge of an L2 can play an important role in TLA provided that it was typologically closer to the target language than the L1 was. Dewaele (2002, as cited in De Angelis, 2007) conducted a study on language anxiety and found that the strongest difference in anxiety levels was between the L1 and the L2.

19 Speaking in the second language causes higher levels of anxiety than speaking in the first language; anxiety then lowers in subsequent languages. This could lead to a conclusion that multilinguals develop their affective strategies to a greater extent than monolinguals, which helps them cope with negative feelings during language production. In spite of a body of evidence suggesting coginitive advantages of bilinguals in comparison to monolinguals when learning L3, it still remains unclear whether or not multilingual learners with more languages would use strategies differently from their peers who studied fewer languages. Even though the studies listed here show that knowing more languages leads to a more diverse and frequent language strategy use, this area of TLA calls for further research. 6 Study 6.1 Aims The aim of this study was to investigate the frequency and potential differences in the use of particular strategies by two groups of Croatian learners; those who studied two foreign languages and those who studied three foreign languages. 6.2 Participants A total of 42 students participated in the research study. Twenty-three students of German as their second foreign language (Group A) and 19 students of German as the second and Italian as the third foreign language (Group B). Students from both groups had English as their first foreign language. Students from the Group A are 18 years old high-school graduates, while those from the Group B are 17 years old 3rd grade students. Students from both groups attend Tin Ujević high school in Kutina. All the participants in the Group A (L3 learners) had Croatian as their mother tongue. The mean of years of learning English is 11.91, while the mean for German is 5.43 years. All the participants in the Group B (L4 learners) also had Croatian as their mother tongue. On average, they had been learning English for 12.05, German for 7.05, and Italian for 3.26 years. Figure 1 shows the students' self-assessment of proficiency in the languages they are learning. The average grades the students in the Group A would give themselves in English

20 and German are 4.21 and 3.56, respectively. On the other hand, the Group B assessed themselves in the following way: 4.57 in English, 3.26 in German, and 2.63 in Italian. 5 4,5 4 3,5 3 2,5 2 1,5 1 0,5 0 Group A (L3 learners) Group B (L4 learners) English German Italian Figure 1: Self-assessment of proficiency in foreign languages, Groups A and B 4,85 4,8 4,75 4,7 4,65 4,6 4,55 4,5 4,45 4,4 4,35 4,3 Group A (L3 learners) English German Italian Group B (L4 learners) Figure 2: Current grades in foreign languages, Groups A and B As can be seen in Figure 2, there is not much difference between the two groups in terms of school grades. Groups A and B had almost the same grades in English (means 4.52 and 4.47 for Groups A and B, respectively) and German (means 4.47 and 4.55).

21 The Group B, who had been studying an additional foreign language, Italian, had an average grade of 4.78 in that language. That contrasts with the self-assessment of proficiency mean, which was 2.63. They had been learning Italian for, on average, only 3.26 years, which, contrasted to a much longer period of time of learning English and German, might have led them to perceive their proficiency as lower. There was also a discrepancy between their selfassesment of proficiency in German and the school grade. This could be explained by the increased awareness of their language performance and the need for improvement. Next, none of the participants in the Group A stated that proficiency in either English or German was unimportant to them. Ninety-one per cent claimed proficiency in English was ''very important'', as opposed to 30 per cent of them stating that for German. In the Group B all the participants reported that proficiency in English was very important to them. 52 per cent of them stated that for German, while only two of them claimed it was ''not important''. Most of them, 84 per cent, agreed that it was ''important'' to be proficient in Italian, with only three of them finding it very important. The participants agreed that English was the most important language to achieve high proficiency in, while German and Italian were considered to be not as important, although the Group B found German slightly more important than the Group A. The Group B chose Italian as an elective in the first grade of high school, so maybe this reflected their general belief about the importance of knowing foreign languages. In the Group A 60 per cent of students stated they enjoyed foreign language learning, while in the Group B only one participant stated he did not, which leads to a conclusion that the Group B were more motivated for learning languages. We can assume that students who are interested in languages and seem to be good at learning them would take an additional language as a school subject, as it was the case in apparently more motivated Group B. 6.3 Data collection instrument Data were collected through a questionnaire consisting of three parts (Appendix 1). The first part elicited background data on the participants' mother tongue, the number of languages they had studied and the length of study. They were also asked to rate their proficiency and state how important they thought high proficiency in the specific foreign languages was. The second part consisted of the statements that were largely based on items from The Strategy Inventory for Language Learning (SILL), which is a self-report questionnaire designed by Oxford (1990) with the aim to assess the frequency of use of language learning strategies.

22 SILL items are evaluated on a five-point Likert scale ranging from 1 to 5, with 1 being ''never or almost never true of me'' and 5 ''always true of me''. The statements represent a wide variety of strategies, from compensation and cognitive to affective and social strategies. Items 7, 12, 23 and 30 elicited cognitive strategies, items 8, 24 and 25 metacognitive strategies, items 19, 20 and 22 compensation strategies, items 26 and 27 affective strategies, items 13 and 28 social strategies, while items 1, 3, 4 and 10 elicited memory strategies. All of these were taken from Oxford (1990). The additional items which tested cross-linguistic strategies were added as well, in order to examine the use of knowledge of other foreign languages in learning an additional one more closely. Those were items: 2, 5, 6, 9, 11, 14, 15, 16, 17, 18, 21 and 31. The third part asked the participants to answer a few open-ended questions dealing with strategy use in learning foreign languages. The aim was to elicit a more detailed response and to see if the participants reported using some strategies that were not mentioned in the second part of the questionnaire. 6.4 Data collection procedures Questionnaires were administered to all the participants on the 4th of May, 2015 by their teachers in Tin Ujević high school in Kutina. Teachers provided brief instructions and explained to the participants what was expected from them, emphasizing that their responses would be completely anonymous. They were instructed to first fill out the background questionnaire, followed by the second part that investigated the frequency of use of learning strategies and the third part which consisted of open-ended questions. StataCorp (2009), a professional statistical software, was used to analyze the collected data. Means were calculated in order to investigate the background information and use of language learning strategies among different groups. 6.5 Results and discussion 6.5.1 Language learning strategies questionnaire analysis The first aim of this research was to investigate the frequency of use of language learning strategies within two groups of Croatian students studying different foreign languages as L3 and L4. The results shown in Figure 5 below represent the arithmetic means of items from SILL, additional items which tested cross-linguistic strategies and overall strategy use calculated separately for two groups of students.

23 4 3,5 3 2,5 2 1,5 1 0,5 0 Items from SILL questionnaire Additional items (cross-linguistic strategies) Overall strategy use Group A (L3 learners) Group B (L4 learners) Figure 5: Language learning strategies questionnaire analysis If we look at the results separately for each group, the Group B (L4 learners) used language learning strategies more often than the Group A (means 3.63 as opposed to 3.07). This may suggest that knowing more languages is reflected in a higher frequency of strategies used when learning an additional language, which is in line with some previous studies, such as the ones by Kemp (2007), Psaltou-Joycey and Kantaridou (2009) and Sung (2011). Mitits and Sarafianou (2012) and Qasimnejad and Hemmati (2013) came to the same conclusions regarding monolinguals and bilinguals, which could be compared with the case of L3 and L4 learners, in that in both cases there is a difference in prior language learning experience leading to a more frequent strategy use. Furthermore, this finding could also mean that the participants who know more languages are more metalingustically aware, as Kemp (2001) suggested, and more conscious of their strategy use in particular situations, while those with fewer languages lack that awareness and are less able to notice when they actually employ strategies, even when they are put to use. As Bialystok (2001, as cited in Kuo & Anderson, 2008) put it, metalinguistic awareness also refers to the attentional control of mental mechanism that operates language processing. In the same vein, Jessner (2006) claimed that, as an emergent property of multilingual proficiency, metalinguistic awareness provided the learners the ability to focus attention on language as an object, which could also refer to the awareness of specific learning techniques.

24 The Group B had a higher mean in items that were taken from the SILL questionnaire, as well as in additional items which tested cross-linguistic strategies. The means for the Group B in each of the groups of items are 3.71 and 3.42. The means for the Group A for the same categories are 3.12 and 2.89. The results have shown that the students used crosslinguistic strategies less often than strategies that items from SILL elicited (means 3.41 and 3.12). 6.5.1.1 Analysis of the items taken from Oxford's SILL (1990) Oxford's items from SILL investigate different categories of strategies, from cognitive to social. In Figures 6 and 7, scores for each of the strategy categories are ranked in the order from the most to the least used. As can be seen from Figures 6 and 7, social (SOC) and metacognitive strategies (MET) are the most often used in both groups, and memory strategies (MEM) are the least often used group of strategies. In the Group A, other strategies are ranked as follows, from the most frequently to the least frequently used: affective (AFF), compensation (COM) and cognitive strategies (COG). In the Group B, ranking of other strategies is: compensation (COM), cognitive (COG) and affective strategies (AFF). Based on the results, we could say that both groups usually use social strategies. In the Group A, almost every other category of strategies could be interpreted as being ''sometimes used''. On the other hand, almost every other category of strategies was ''usually used'' by the Group B. However, the exception in both groups were memory strategies, which could be interpreted as being ''rarely used'' in the Group A, and as being ''sometimes used'' in the Group B. Figure 6: The means for strategy categories of items taken from SILL; Group A

25 Figure 7: The means for strategy categories of items taken from SILL; Group B The results of Korkmaz's (2013) and Qasimnejad and Hemmati's (2013) studies, which found that bilingual university students used social strategies the least often, are not in line with these results. The reason for social strategies' high score in this study might be the fact that secondary education curriculum is based on a lot of team work, research and projects. High-school students are more encouraged than in elementary education to work together, not only in foreign language classes, but in other subjects as well. They may also be more interested in spending time with their peers and exploring the world. Moreover, as they grow older and gain more knowledge on language learning, they start realizing that in order to learn a language well, it needs to be practiced as often as possible. Furthermore, Thomas (1982) found that bilinguals used more communication strategies and attributed this to their need to switch languages, which could enhance their sensitivity to the functions of language for social communication. All of that might contribute to social strategies taking the first position. Metacognitive strategies are the second most often used strategies in both groups. This is in line with the studies by Qasimnejad and Hemmati (2013) and Sung (2011), who found this category to be among the most often used strategy categories. They found that bilinguals control their progress in foreign language learning and seek out ways to improve their learning. As a result of learning more than one foreign language over the years, the participants in this study had gained knowledge on language learning process that helps them plan for, monitor and evaluate the success of a learning activity. They are also more